Statistical method of analyzing long-term tropospheric loss

A detailed exposition of multiple correlation applied to the determination of the relationship between signal strength variations from day to day and weather observations. Adoption of a screening procedure using a computer programme permits the more significant weather variables to bo determined. Data used was obtained from various Mediterranean sea paths and the largely overland path London-Paris. The deduced relationship formula giving the best correlation for each signal circuit differs considerably from those for other circuits, but it is suggested that this may be due partly to their different geographical/weather pattern characteristics and partly to insufficiency of data. Some improvement in the mathematical model may be possible by weighting the data for days of low signal strength variance (which are generally days of low average signal strength), since the data for these days is more precise and the propagation mode generally more stable. The statistical techniques and assumptions made are discussed in some detail and 9 references given. Sample results are tabulated, including illustrations of the step-by-step screening procedure.